There are two main types of integrated detector in existence, namely charge-coupled devices (CCD) in which the image is stored and then transferred pixel by pixel to an acquisition system, and active pixel detectors (APD) in which each pixel is biased and preamplified separately.
Some detectors, in particular those using superconducting tunnel junctions (STJs) can be used only in APD mode. An absorbed photon or particle generates a signal representing excess charge or current that must be read immediately, in that it cannot be stored for multiplexing purposes as is possible with CCD type sensors. This means that an individual acquisition system is required for each pixel. This in turn implies that numerous circuits must be implemented, and above all that it can become difficult or even impossible to implement the necessary connections when the number of pixels is high. The problem becomes even more critical with cryogenic detectors (e.g. STJ type detectors) where each additional connection wire increases heat flow into the cryogenic portion.
European patent application EP 0 809 394 (Kodak) proposes an architecture which enables read systems to be simplified, but without addressing the need to associate one such system with each pixel.
An object of the invention is to provide a detector which enables the number of read systems and the number of connections to be decreased.
The invention thus provides an active pixel type detector comprising M rows and N columns, each pixel having a top detector electrode and a bottom detector electrode, wherein the top detector electrodes are interconnected in rows or columns and the bottom electrodes are interconnected in columns or rows to form a row interconnection output for each row and to form a column interconnection output for each column, each row being taken to a first potential and being coupled to a row amplifier, and each column being taken to a second potential and being coupled to a column amplifier.
Whereas known detectors require Mxc3x97N+1 connections and Mxc3x97N read systems to be implemented, the device of the invention requires only M+N connections and M+N read systems. In particular, for a square matrix, having N2 pixels (M=N), the device of the invention requires only 2N connections, and 2N read systems, instead of N2+1 connections and N2 read systems as is required in the prior art.
The row amplifiers may be high gain amplifiers, each having a first input taken to the first potential and having a second input connected to a corresponding one of said row interconnection outputs.
The column amplifiers may be high gain amplifiers each having a first input taken to the second potential and a second input connected to a corresponding one of said column interconnection outputs.
The invention applies in particular to superconducting tunnel junction detectors.